From 93562ca60740f8a492eb0f7038afb967263de6b4 Mon Sep 17 00:00:00 2001 From: Riku-Laine <28960190+Riku-Laine@users.noreply.github.com> Date: Mon, 10 Jun 2019 13:16:43 +0300 Subject: [PATCH] Modifications and writing --- Kandi.tex | 85 ++++-- .../Analysis_07MAY2019_new.ipynb | 269 +++++++++++++++--- 2 files changed, 298 insertions(+), 56 deletions(-) diff --git a/Kandi.tex b/Kandi.tex index fd08d93..8f0b765 100644 --- a/Kandi.tex +++ b/Kandi.tex @@ -9,6 +9,10 @@ \usepackage{amsmath} \usepackage{amssymb} +\usepackage{pgf} +\usepackage{tikz} +\usetikzlibrary{arrows,automata} + \usepackage{hyperref} \usepackage{url} \usepackage{icomma} @@ -159,12 +163,11 @@ Tämän tutkielman tavoitteena on luoda kausaalipäättelyn avulla algoritmi, jo % miksi halutaan siirtyä (frekventistisen / bayes-päättelyn ongelmat), edut, esiintyminen, erot, käyttö -Kuten Pearl ja Mackenzie esittävät kirjassaan Miksi, ihmisillä on luontainen kausaalisen päättelyn taito \cite{miksi}. Tavalliset tilastollisen päättelyn menetelmät eivät tarjoa tapaa määritellä kausaalista yhteyttä: aineistosta voidaan päätellä erilaisia \emph{korrelaatioita}, mutta kausaalinen päättely \emph{A johtuu B:stä} vaatii uudenlaista lähestymistapaa. Käytännön tutkimuksessa kausaaliset yhteydet kiinnostavat erityisesti lääketieteen alalla \cite{pearl10}. Kuten Kalisch toteaa, aiemmin kausaalisuuden päättely on perustunut korrelaatioiden havaitsemiseen. On hypotetisoitu, että biomarkkerin ja taudin samanaikainen ilmaantuminen viittaisi siihen, että markkeri aiheuttaa taudin. Voimmeko siis markkeria käsittelemällä vaikuttaa tautiin tai jopa parantaa se? \cite{kalisch14} - -Syy-seuraussuhteen matemaattinen määrittely vaatii uutta lähestymistä myös todennäköisyyslaskennan merkintöihin. Pearl käyttää alkuperäisessä, englanninkielisessä kirjallisuudessa merkintää 'do' ilmaisemaan interventiota. Merkinnällä halutaan erottaa tavanomainen ehdollinen todennäköisyys $\pr(Y|X=x)$ interventiosta, jossa asetamme muuttujan $X$ arvoon $x$: $\pr(Y|\text{do}(X=x))$. Kimmo Pietiläinen käyttää kirjan suomennoksessa do-operaattorista käännöstä \emph{tee}, mutta seuraan tässä tutkielmassa Pearlin merkintöjä, ellen erikseen muuta mainitse \cite{miksi}. Alalla käytetään myös muita, alaindekseillä rikastettuja merkintätapoja \cite{pearl10}. Esittelen käyttämäni merkinnät tarkemmin kappaleessa \ref{kausaalimerk_laus}. +Judea Pearl ja Mackenzie esittävät kirjassaan Miksi, että ihmisillä on luontainen kausaalisen päättelyn taito \cite{miksi}. Tavalliset tilastollisen päättelyn menetelmät eivät tarjoa tapaa määritellä kausaalista yhteyttä: aineistosta voidaan päätellä erilaisia \emph{korrelaatioita}, mutta kausaalista päättelyä \emph{A johtuu B:stä} ei voida tehdä perinteisen tilastotieteen keinoin. Käytännön tutkimuksessa kausaaliset yhteydet kiinnostavat erityisesti lääketieteen alalla \cite{pearl10}. Kuten Kalisch toteaa, aiemmin kausaalisuuden päättely on perustunut korrelaatioiden havaitsemiseen. On hypotetisoitu, että jonkinlaisen biomarkkerin ja taudin samanaikainen havaitseminen viittaisi siihen, että markkeri aiheuttaa taudin. Voimmeko siis markkeria käsittelemällä vaikuttaa tautiin tai jopa parantaa sen? \cite{kalisch14} -Kausaalipäättelyssä mallit voidaan esittää graafeina, eli verkkoina. Verkoista voidaan suoraan lukea eri muuttujien relaatiot kausaalisuuden suuntien ja riippuvuuksien suhteen. +Syy-seuraussuhteen matemaattinen määrittely vaatii uutta lähestymistä myös todennäköisyyslaskennan merkintöihin. Pearl käyttää alkuperäisessä, englanninkielisessä kirjallisuudessa merkintää 'do' ilmaisemaan interventiota. Merkinnällä halutaan erottaa tavanomainen ehdollinen todennäköisyys $\pr(Y|X=x)$ interventiosta, jossa asetamme muuttujan $X$ arvoon $x$: $\pr(Y|\text{do}(X=x))$. Kausaalipäättelyyn liittyvät myös oleellisesti kontrafaktuaalit, jotka kuvaavat muuttujien mahdollisia arvoja, jos jokin toinen muuttuja olisi ollut erilainen -- "sateen todennäköisyys, jos taivaalla olisi ollut pilviä". Esimerkiksi muuttujan $Y$ arvoa, jos $X$ olisi ollut $x$ asteikolla $u$ merkittäisiin $Y_x(u)$. Tässä tutkielmassa käsittelen kuitenkin vain Pearlin kausaalimallia. Esittelen käyttämäni merkinnät tarkemmin kappaleessa \ref{kausaalimerk_laus}. +Kausaalipäättelyssä mallit voidaan esittää graafeina, eli verkkoina. Verkoista voidaan suoraan lukea eri muuttujien syy-seuraussuhteet ja riippuvuudet tai riippumattomuudet. %%%%%%%%% \section{Valikoitumisharha -- seulotun aineiston ongelma}\label{sl} @@ -175,7 +178,7 @@ Tässä tutkielmassa tarkasteltavassa asetelmassa havaintojen puuttuminen liitty Aineiston luova mekanismi on esitetty kuvassa \ref{valikoitumisharha} ja toimii siten, että aluksi jokin henkilö tai muu entiteetti saapuu päätöksentekijän eteen seulottavaksi. Päätöksentekijän tavoitteena on estää haitallinen tulos ($y=0$) pitäen samalla myönteisten päätösten ($t=1$) määrä mahdollisimman pienenä. Seuloja pyrkii siis antamaan kielteisen päätöksen kaikille niille, joilla epäonnistuminen on todennäköisin. Päätöksen jälkeen Kohtalo määrittää havainnolle tuloksen $y\in\{0, 1\}$. Kielteisen päätöksen saaneille tulos voidaan merkitä puuttuvaksi tai onnistuneeksi, koska haitallista tapahtumaa ei havaita. -Aineiston generoivaa mekanismia voidaan havainnollistaa lääke- ja oikeustieteen alan esimerkeillä. Henkilö on ensin mainitussa potilas ja jälkimmäisessä epäilty. Seuloja voi olla esimerkiksi lääkäri, joka päättää annetaanko potilaalle vahvempaa ja samalla kalliimpaa lääkettä, jolloin relapsia ei havaita. Oikeudellisessa asetelmassa seulojalla voidaan tarkoittaa tuomaria, joka päättää epäillyn vapauttamisesta takuita vastaan ilman pelkoa rikoksen uusimisesta. Molemmilla päättäjillä on selkeä kannustin estää haitalliset tulokset -- sairauskohtaukset tai rikokset -- pitäen samalla päätöksistä aiheutuvat rasitteet yhteiskunnalle ja yksilöiden elämille mahdollisimman pienenä. Lisäksi erityisesti oikeudellisessa asetelmassa on selvää, kuinka takuukäsittelystä kielteisen tuloksen saaneet eivät voi syyllistyä uuteen rikokseen, joten heidän tulosmuuttujan arvo voidaan koodata joko onnistumiseksi tai havaitsemattomaksi. +Aineiston generoivaa mekanismia voidaan havainnollistaa lääke- ja oikeustieteen alan esimerkeillä. Henkilö on ensin mainitussa potilas ja jälkimmäisessä epäilty. Seuloja voi olla esimerkiksi lääkäri, joka päättää annetaanko potilaalle vahvempaa ja samalla kalliimpaa lääkettä, jolloin relapsia ei havaita. Oikeudellisessa asetelmassa seulojalla voidaan tarkoittaa tuomaria, joka päättää epäillyn vapauttamisesta takuita vastaan ilman pelkoa rikoksen uusimisesta. Molemmilla päättäjillä on selkeä kannustin estää haitalliset tulokset -- sairauskohtaukset tai rikokset -- pitäen samalla päätöksistä aiheutuvat rasitteet yhteiskunnalle ja yksilöiden elämille mahdollisimman pienenä. Lisäksi erityisesti oikeudellisessa asetelmassa on selvää, kuinka takuukäsittelystä kielteisen tuloksen saaneet eivät voi syyllistyä uuteen rikokseen, joten heidän tulosmuuttujan arvo voidaan merkitä joko onnistumiseksi tai havaitsemattomaksi. \begin{figure}%[H] \centering @@ -230,7 +233,7 @@ Aineistossa jyvitämme jokaiselle $M=100$ päätöksentekijälle 500 arvioitavaa \end{equation} mukaisesti. Jos $\pr(Y=0|X, Z, W) \geq 0,5$, tulosmuuttujan arvoksi asetetaan 0 ja vastaavasti jos $\pr(Y=0|X, Z, W) < 0,5$ muuttujan arvoksi asetetaan 1. Lausekkeissa \ref{y_ehd} ja \ref{t_ehd} olevat kertoimet $\beta_X$, $\beta_Z$ ja $\beta_W$ ovat 1, 1 ja 0,2 vastaavassa järjestyksessä. \cite{lakkaraju17} -Päätösmuuttuja $T$ määritetään kaksivaiheisesti: ensin määritetään todennäköisyys kielteiselle päätökselle ja sitten muuttujan arvo asetetaan näiden todennäköisyyksien keskinäisen suuruuden mukaisesti. Muuttujan $T$ ehdollinen todennäköisyys +Päätösmuuttuja $T$ määritetään kaksivaiheisesti: ensin määritetään todennäköisyys kielteiselle päätökselle ja sitten muuttujan arvo asetetaan näiden todennäköisyyksien keskinäisen suuruuden ja hyväksymisasteen $r$ mukaisesti. Muuttujan $T$ ehdollinen todennäköisyys \begin{equation} \label{t_ehd} \pr(T=0|X, Z)=\frac{1}{1+\text{exp}\{-(\beta_XX+\beta_ZZ)\}} + \epsilon, \end{equation} @@ -265,33 +268,55 @@ Muuttuja & Keskiarvo & Keskihajonta & Minimi & 25\% & 50\% & Tässä kappaleessa esitän tutkielmassani käyttämät metodit. Selostan supistusalgoritmin toiminnan kappaleessa \ref{contraction} sekä kausaalisen mallin laatimisessa ja arvioinnissa käyttämäni teoreettisen taustan kappaleissa \ref{verkot}. Koska kausaalinen malli esitetään verkkona, käyn aluksi läpi vaadittavat verkkoteoreettiset määritelmät. Esitän sen jälkeen mallini graafina ja osoitan kausaalisen vaikutuksen olevan identifioituva. +\section{Metriikat} + +Algoritmien suorituskyvyn arviointiin liittyy kolme keskeistä metriikkaa: hyväksymisaste (engl. \emph{acceptance rate}, (AR)), epäonnistumisprosentti (\emph{failure rate} (FR)) ja keskimääräinen virhe(\emph{mean absolute error} (MAE)). + +\begin{maar}[Hyväksymisaste (AR)] \label{acc_rate} + +Päättäjän hyväksymisaste määritetään myönteisten päätösten määrän suhteena annettujen päätösten kokonaismäärään. Eli jos päätöksentekijä antaa 100 päätöstä, joista 40 on myönteisiä, niin hänen hyväksymisasteensa on $0,4$. + +\end{maar} + +\begin{maar}[Epäonnistumisprosentti (FR)] \label{fail_rate} + +Päätöksentekijän epäonnistumisprosentti määritetään epäonnistuneiden tulosten määrän suhteena annettujen päätösten kokonaismäärään. Eli jos päätöksentekijä antaa 100 päätöstä, joista 60 on myönteistä ja näistä 60 päätöksestä 30 johtaa epäonnistumiseen (esimerkiksi rikoksen uusintaan), niin tuomarin epäonnistumisprosentti on $0,3$. + +\end{maar} + +\begin{maar}[Keskimääräinen virhe (MAE)] \label{mae} + + + +\end{maar} + \section{Supistusalgoritmi}\label{contraction} -Supistusalgoritmin on lakkarajun .... +Supistusalgoritmi on 2017 esitetty algoritmi \cite{lakkaraju17}, jonka avulla voidaan arvioida ennustavien mallien todellista suorituskykyä, kun vain tietylle osalle aineistosta on luokka (label) saatavissa. Algoritmin toimintaperiaatteena on arvioida mallin $\B$ ennusteita löyhimmän päätöksentekijän tekemien päätösten joukossa. Algoritmin pseudokoodi on esitetty Algoritmissa \ref{contraction_alg}. \begin{algorithm}[H] % enter the algorithm environment \caption{Supistusalgoritmi} % give the algorithm a caption \label{contraction_alg} % and a label for \ref{} commands later in the document \begin{algorithmic}[1] % enter the algorithmic environment \REQUIRE Aineisto $\D$, todennäköisyydet $\s$ ja hyväksymisaste $r$ -\ENSURE failure rate at acceptance rate r +\ENSURE Epäonnistumisprosentti (FR) hyväksymisasteella $r$ -\STATE Olkoon $q$ päättäjä, jolla on korkein $r$. +\STATE Olkoon $q$ päättäjä, jolla on korkein hyväksymisaste $r$. \STATE $\D_q = \{(x, j, t, y) \in \D | j = q \}$ -\STATE Nyt $\D_q$ on havaintojoukko, jolle $q$ on antanut päätökset. +\STATE \hskip3.0em $\rhd$ Nyt $\D_q$ on havaintojoukko, jolle $q$ on antanut päätökset. \STATE \STATE $\RR_q = \{(x, j, t, y) \in \D_q | t=1 \}$ -\STATE $\RR_q$ on on joukon $\D_q$ osa, jolle tulosmuuttujan arvot on havaittu. +\STATE \hskip3.0em $\rhd$ $\RR_q$ on on joukon $\D_q$ osa, jolle tulosmuuttujan arvot on havaittu. \STATE -\STATE Järjestä aineiston $\RR_q$ havainnot laskevaan järjestykseen todennäköisyyksien $\s$ mukaan ja talleta taulukkoon $\RR_q^{sort}$ -\STATE Mallin korkeariskisimmät ovat nyt listan kärjessä +\STATE Järjestä taulukoon $\RR_q$ havainnot laskevaan järjestykseen todennäköisyyksien $\s$ mukaan ja talleta ne taulukkoon $\RR_q^{sort}$ +\STATE \hskip3.0em $\rhd$ Mallin korkeariskisimmät ovat nyt listan kärjessä \STATE -\STATE Ota aineistosta $\RR_q^{sort}$ sen $[(1.0-r)|\D_q|]-[|\D_q|-|\RR_q|]$ ensimmäistä/ylintä havaintoa ja talleta ne taulukkoon $\RR_\B$. -\STATE $\RR_\B$ on mallin b käsittää ne henkilöt, jolle malli $\B$on antanut positiivisen päätöksen +\STATE Ota taulukosta $\RR_q^{sort}$ sen $[(1.0-r)|\D_q|]-[|\D_q|-|\RR_q|]$ ensimmäistä/ylintä havaintoa ja talleta ne taulukkoon $\RR_\B$. +\STATE \hskip3.0em $\rhd$ $\RR_\B$ on lista henkilöistä, joille malli $\B$ on antanut positiivisen päätöksen \STATE -\STATE Laske $u = \sum_{i=1}^{|\RR_\B|} \dfrac{\delta\{y_i=0\}}{|\D_q|}$ +\STATE Laske $\mathbf{u} = \sum_{i=1}^{|\RR_\B|} \dfrac{\delta\{y_i=0\}}{|\D_q|}$. -\RETURN $u$ +\RETURN $\mathbf{u}$ \end{algorithmic} \end{algorithm} @@ -317,14 +342,34 @@ Verkot koostuvat \emph{solmuista} ja \emph{kaarista}, joita voidaan havainnollis % TiRan materiaalit?? % Ota esimerkki verkko ja kirjoita siitä lyhyet havainnollistavat kommentit - -\begin{figure}[H] + +\begin{figure} [h] \centering -\includegraphics[scale = 0.5]{full_model} +\begin{tikzpicture}[->,>=stealth',node distance=2.0cm, semithick] + + \tikzstyle{every state}=[fill=none,draw=black,text=black] + + \node[state] (R) {$R$}; + \node[state] (X) [right of=R] {$X$}; + \node[state] (T) [below of=X] {$T$}; + \node[state] (Z) [right of=X] {$Z$}; + \node[state] (Y) [below of=Z] {$Y$}; + + \path (R) edge (T) + (X) edge (R) + edge (T) + edge (Y) + (Z) edge (X) + edge (T) + edge (Y) + edge [bend right] (R) + (T) edge (Y); +\end{tikzpicture} \caption{Eräs verkko $H = (V, E)$, missä $V = \{R, X, Z, T, Y\}$.} \label{esverkko} \end{figure} + \begin{maar}[Suunnattu verkko] \label{suun_verkko} \emph{Suunnattu verkko G} on pari $(V, E)$, missä $V \neq \emptyset$ on \emph{solmujen} joukko ja $$E = \{(a, b) \in V \times V | \text{ solmusta } a \text{ on nuoli solmuun } b \} $$ on \emph{kaarien} joukko. diff --git a/analysis_and_scripts/Analysis_07MAY2019_new.ipynb b/analysis_and_scripts/Analysis_07MAY2019_new.ipynb index b8d8ae9..0007648 100644 --- a/analysis_and_scripts/Analysis_07MAY2019_new.ipynb +++ b/analysis_and_scripts/Analysis_07MAY2019_new.ipynb @@ -56,7 +56,7 @@ }, { "cell_type": "code", - "execution_count": 1, + "execution_count": 51, "metadata": {}, "outputs": [], "source": [ @@ -79,8 +79,8 @@ "\n", "%matplotlib inline\n", "\n", - "plt.rcParams.update({'font.size': 20})\n", - "plt.rcParams.update({'figure.figsize': (14, 7)})\n", + "# plt.rcParams.update({'font.size': 20})\n", + "# plt.rcParams.update({'figure.figsize': (14, 7)})\n", "\n", "# Suppress deprecation warnings.\n", "\n", @@ -124,7 +124,7 @@ }, { "cell_type": "code", - "execution_count": 3, + "execution_count": 52, "metadata": {}, "outputs": [], "source": [ @@ -221,7 +221,7 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 53, "metadata": {}, "outputs": [], "source": [ @@ -311,7 +311,7 @@ }, { "cell_type": "code", - "execution_count": 4, + "execution_count": 54, "metadata": {}, "outputs": [], "source": [ @@ -415,7 +415,7 @@ }, { "cell_type": "code", - "execution_count": 5, + "execution_count": 55, "metadata": {}, "outputs": [], "source": [ @@ -465,7 +465,7 @@ "\n", " prediction_x_0 = prediction(x_0)\n", "\n", - " x_values = np.linspace(-10, 10, 40000)\n", + " x_values = np.linspace(-15, 15, 40000)\n", "\n", " x_preds = prediction(x_values)\n", "\n", @@ -605,7 +605,7 @@ }, { "cell_type": "code", - "execution_count": 6, + "execution_count": 56, "metadata": {}, "outputs": [], "source": [ @@ -653,7 +653,7 @@ }, { "cell_type": "code", - "execution_count": 8, + "execution_count": 57, "metadata": { "scrolled": false }, @@ -662,12 +662,12 @@ "name": "stdout", "output_type": "stream", "text": [ - "[1] 0 1 2 3 4 5 6 7 8 9 [2] 0 1 2 3 4 5 6 7 8 9 [3] 0 1 2 3 4 5 6 7 8 9 [4] 0 1 2 3 4 5 6 7 8 9 [5] 0 1 2 3 4 5 6 7 8 9 [6] 0 1 2 3 4 5 6 7 8 9 [7] 0 1 2 3 4 5 6 7 8 9 [8] 0 1 2 3 4 5 6 7 8 9 " + "[1] 0 1 [2] 0 1 [3] 0 1 [4] 0 1 [5] 0 1 [6] 0 1 [7] 0 1 [8] 0 1 " ] }, { "data": { - "image/png": "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\n", + "image/png": "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\n", "text/plain": [ "<Figure size 1008x576 with 1 Axes>" ] @@ -681,20 +681,20 @@ "name": "stdout", "output_type": "stream", "text": [ - "[[0.015368 0.005808 0.01995777 0.01516406 0.01594067]\n", - " [0.041416 0.015836 0.04321981 0.04789159 0.04243067]\n", - " [0.076044 0.0261 0.07418454 0.07789905 0.07412071]\n", - " [0.116112 0.039852 0.11635393 0.11883984 0.11509014]\n", - " [0.162124 0.052308 0.16481971 0.16338339 0.16305947]\n", - " [0.216192 0.06732 0.21387284 0.23024922 0.21453262]\n", - " [0.274452 0.080512 0.27114682 0.27166327 0.27669359]\n", - " [0.3408 0.100856 0.34501975 0.35435532 0.34294369]]\n", + "[[0.01526 0.00628 0.01889946 0.01396761 0.01605884]\n", + " [0.04178 0.01518 0.04507529 0.05034328 0.04077152]\n", + " [0.07558 0.02656 0.08102275 0.075079 0.07462003]\n", + " [0.11526 0.0389 0.12120447 0.12251311 0.11627801]\n", + " [0.1648 0.05258 0.1714974 0.16779408 0.16470203]\n", + " [0.21774 0.06782 0.20684378 0.23026316 0.21417139]\n", + " [0.27726 0.08108 0.27048099 0.29930628 0.27578153]\n", + " [0.3431 0.08988 0.35903947 0.35946718 0.34240791]]\n", "\n", "Mean absolute errors:\n", - "0.1067395\n", - "0.002629345839201819\n", - "0.005365384751435717\n", - "0.001439077136508043\n" + "0.10906249999999998\n", + "0.007329259700714057\n", + "0.008942559709348354\n", + "0.0012028047169419824\n" ] } ], @@ -702,7 +702,9 @@ "f_rates = np.zeros((8, 5))\n", "f_sems = np.zeros((8, 5))\n", "\n", - "nIter = 10\n", + "nIter = 2\n", + "\n", + "#npr.seed(0)\n", "\n", "for r in np.arange(1, 9):\n", "\n", @@ -733,7 +735,7 @@ "\n", " #### True evaluation\n", " # Sort by actual failure probabilities, subjects with the smallest risk are first.\n", - " s_sorted = s_test.sort_values(by='probabilities_Y',\n", + " s_sorted = s_test.sort_values(by='B_prob_0_logreg',\n", " inplace=False,\n", " ascending=True)\n", "\n", @@ -780,9 +782,9 @@ " r / 10)\n", " #### Causal model\n", "\n", - " released = bailIndicator(r * 10, s_logreg, s_train.X, s_test.X)\n", + " #released = bailIndicator(r * 10, s_logreg, s_train.X, s_test.X)\n", "\n", - " #released = cdf(s_test.X, s_logreg, 0) < r/10\n", + " released = cdf(s_test.X, s_logreg, 0) < r/10\n", "\n", " s_f_rate_caus[i] = np.mean(s_test.B_prob_0_logreg * released)\n", "\n", @@ -873,14 +875,56 @@ }, { "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], + "execution_count": 74, + "metadata": { + "scrolled": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "[1] 0 1 [2] 0 1 [3] 0 1 [4] 0 1 [5] 0 1 [6] 0 1 [7] 0 1 [8] 0 1 " + ] + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "<Figure size 1008x576 with 1 Axes>" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "[[0.0055 0.00276 0.01342565 0.00618992 0.00381115]\n", + " [0.02222 0.00974 0.02697824 0.02987324 0.01296137]\n", + " [0.04612 0.01816 0.05032907 0.02793117 0.02842223]\n", + " [0.08052 0.0301 0.08272431 0.0629776 0.0499887 ]\n", + " [0.12892 0.04876 0.12535567 0.11325511 0.08063283]\n", + " [0.18022 0.0635 0.18180218 0.15531875 0.11735564]\n", + " [0.24572 0.07712 0.25421899 0.22589487 0.16578486]\n", + " [0.31992 0.1095 0.31794241 0.30081773 0.23036298]]\n", + "\n", + "Mean absolute errors:\n", + "0.0836875\n", + "0.004340044672262239\n", + "0.015445989730460349\n", + "0.04247753106139743\n" + ] + } + ], "source": [ "failure_rates = np.zeros((8, 5))\n", "failure_sems = np.zeros((8, 5))\n", "\n", - "nIter = 5\n", + "nIter = 2\n", "\n", "for r in np.arange(1, 9):\n", "\n", @@ -913,6 +957,53 @@ "\n", " test_labeled = test_labeled.assign(B_prob_0_logreg=predictions_labeled)\n", "\n", + "# # Regress T on X\n", + "# lr_t, __ = fitLogisticRegression(train_labeled.X,\n", + "# train_labeled.decision_T, np.ones(1),\n", + "# 1)\n", + "# # Calculate the residuals from previous regression\n", + "# residuals_T = train_labeled.decision_T - \\\n", + "# lr_t.predict(train_labeled.X.values.reshape(-1, 1))\n", + "# train_labeled = train_labeled.assign(residuals_T=residuals_T)\n", + "\n", + "# # Convert residuals from -1, 0 and 1 values to one-hot-encoded.\n", + "# # this way there will be separate betas for each type of residual.\n", + "# enc = OneHotEncoder(categories='auto', drop='first')\n", + "# resid_tf = train_labeled.residuals_T.values.reshape(-1, 1)\n", + "# tmp = enc.fit_transform(resid_tf).toarray()\n", + "# train_labeled = train_labeled.assign(residuals_1=tmp[:, 0],\n", + "# residuals_2=tmp[:, 1])\n", + "\n", + "# # Regress Y on X and residuals from step 2.\n", + "# lr_y, __ = fitLogisticRegression(\n", + "# train_labeled.dropna()[['X', 'residuals_1', 'residuals_2']],\n", + "# train_labeled.dropna().result_Y, np.ones((1, 3)), 0)\n", + "# # With the test data, predict Y by\n", + "# # repeating steps 1 and 2\n", + "# # (Regress T on X)\n", + "# lr_t, __ = fitLogisticRegression(test.X,\n", + "# test.decision_T, np.ones(1),\n", + "# 1)\n", + "\n", + "# # (Calculate the residuals from previous regression)\n", + "# residuals_T = test.decision_T - \\\n", + "# lr_t.predict(test.X.values.reshape(-1, 1))\n", + "# test = test.assign(residuals_T=residuals_T)\n", + "\n", + "# # (Convert residuals from -1, 0 and 1 values to one-hot-encoded.\n", + "# # this way there will be separate betas for each type of residual.)\n", + "# enc = OneHotEncoder(categories='auto', drop='first')\n", + "# resid_tf = test.residuals_T.values.reshape(-1, 1)\n", + "# tmp = enc.fit_transform(resid_tf).toarray()\n", + "# test = test.assign(residuals_1=tmp[:, 0], residuals_2=tmp[:, 1])\n", + "\n", + "# # by using the model from step 3 with X and the residuals from 4.a. as input\n", + "\n", + "# preds = getProbabilityForClass(\n", + "# test[['X', 'residuals_1', 'residuals_2']], lr_y, 0)\n", + "\n", + "# test = test.assign(preds=preds)\n", + "\n", " # True evaluation\n", " #\n", " # Sort by failure probabilities, subjects with the smallest risk are first.\n", @@ -961,18 +1052,20 @@ "\n", " # Causal model - empirical performance\n", "\n", - " released = bailIndicator(r * 10, logreg, train.X, test.X)\n", + "# released = bailIndicator(\n", + "# r * 10, lr_y, train_labeled[['X', 'residuals_1', 'residuals_2']],\n", + "# test[['X', 'residuals_1', 'residuals_2']])\n", "\n", - " #released = cdf(test.X, logreg, 0) < r / 10\n", + " released = cdf(test.X, logreg, 0) < r / 10\n", "\n", " f_rate_caus[i] = np.mean(test.B_prob_0_logreg * released)\n", "\n", " #percentiles = estimatePercentiles(train_labeled.X, logreg, N_sample=train_labeled.shape[0])\n", "\n", - " #def releaseProbability(x):\n", + " # def releaseProbability(x):\n", " # return calcReleaseProbabilities(r*10, train_labeled.X, x, logreg, percentileMatrix=percentiles)\n", "\n", - " #def integraali(x):\n", + " # def integraali(x):\n", " # p_y0 = logreg.predict_proba(x.reshape(-1, 1))[:, 0]\n", "\n", " # p_t1 = releaseProbability(x)\n", @@ -1036,6 +1129,110 @@ "for i in range(1, failure_rates.shape[1]):\n", " print(np.mean(np.abs(failure_rates[:, 0] - failure_rates[:, i])))" ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "scrolled": true + }, + "outputs": [], + "source": [ + "# import pystan\n", + "\n", + "# code = \"\"\"\n", + "# data {\n", + "# int<lower=0> N_obs;\n", + "# int<lower=0> N_mis;\n", + " \n", + "# int y_obs[N_obs];\n", + "# vector[N_obs] x_obs;\n", + "# int t_obs[N_obs];\n", + " \n", + "# vector[N_mis] x_mis;\n", + "# int t_mis[N_mis];\n", + " \n", + "# // int<lower = 1, upper = N_obs + N_mis> ii_obs[N_obs];\n", + "# // int<lower = 1, upper = N_obs + N_mis> ii_mis[N_mis];\n", + "# }\n", + "\n", + "# transformed data {\n", + "# int N = N_mis + N_obs;\n", + "# // vector[N] X;\n", + "# // X[ii_obs] = x_obs;\n", + "# // X[ii_mis] = x_mis;\n", + "# // real[N] T;\n", + "# // T[ii_obs] = t_obs;\n", + "# // T[ii_mis] = t_mis;\n", + "# }\n", + "\n", + "# parameters {\n", + "# real a_;\n", + "# real b_;\n", + "# real c_;\n", + "# real d_;\n", + "# real e_;\n", + "# real f_;\n", + "# vector[N_obs] z_obs;\n", + "# vector[N_mis] z_mis;\n", + "# }\n", + "\n", + "# // transformed parameters {\n", + "# // vector[N] Z;\n", + "# // Z[ii_obs] = z_obs;\n", + "# // Z[ii_mis] = z_mis;\n", + "# // }\n", + "\n", + "# model {\n", + "# z_obs ~ normal(0, 1);\n", + "# z_mis ~ normal(0, 1);\n", + "# //Z ~ normal(0, 1);\n", + "# y_obs ~ bernoulli_logit(d_ * x_obs + e_ * z_obs + f_);\n", + "# t_obs ~ bernoulli_logit(a_ * x_obs + b_ * z_obs + c_);\n", + "# t_mis ~ bernoulli_logit(a_ * x_mis + b_ * z_mis + c_);\n", + "# //t_obs ~ bernoulli_logit(d_ * X + e_ * Z + f_);\n", + "# }\n", + "# \"\"\"\n", + "\n", + "# dat = dict(N_obs = int(train_labeled.dropna().shape[0]/5), \n", + "# N_mis = int(train_labeled[train_labeled.decision_T==0].shape[0]/5),\n", + "# y_obs = train_labeled.dropna().result_Y[::5].astype(int),\n", + "# x_obs = train_labeled.dropna().X[::5],\n", + "# t_obs = train_labeled.dropna().decision_T[::5],\n", + "# x_mis = train_labeled.X[train_labeled.decision_T==0][::5],\n", + "# t_mis = train_labeled.decision_T[train_labeled.decision_T==0][::5])\n", + "\n", + "# sm = pystan.StanModel(model_code=code)\n", + "# fit = sm.sampling(data=dat, iter=8000, chains=4)\n", + "\n", + "import pystan\n", + "\n", + "code = \"\"\"\n", + "data {\n", + " int<lower=0> N_obs; \n", + " int y_obs[N_obs];\n", + " vector[N_obs] x_obs;\n", + "}\n", + "\n", + "parameters {\n", + " real d_;\n", + " real e_;\n", + " vector[N_obs] Z;\n", + "}\n", + "\n", + "model {\n", + " Z ~ normal(0, 1);\n", + " y_obs ~ bernoulli_logit(d_ * x_obs + e_ * Z);\n", + "}\n", + "\"\"\"\n", + "\n", + "# dat = dict(N_obs = int(train_labeled.dropna().shape[0]/18), \n", + "# y_obs = train_labeled.dropna().result_Y[::18].astype(int),\n", + "# x_obs = train_labeled.dropna().X[::18])\n", + "\n", + "# sm = pystan.StanModel(model_code=code)\n", + "# fit = sm.sampling(data=dat, iter=10000, chains=4, control=dict(max_treedepth=17))" + ] } ], "metadata": { -- GitLab